Transcriptome Dynamics Underlying Rod Development: After the final mitosis, morphogenesis and functional differentiation of photoreceptors in the mouse retina span over period of three weeks. Dynamic yet precise changes in gene expression patterns are associated with each developmental stage. We generated transcriptome profiles of purified developing rods by taking advantage of GFP driven by the Nrl promoter (Nrl-GFP). As NRL initiates all the downstream events leading to functional rods, we also produced expression profiles of Nrl-/-:Nrl-GFP photoreceptors that exhibit S-cone characteristics due to the loss of Nrl. We have integrated the temporal transcriptome datasets to NRL ChIP-seq data. Our studies demonstrated a major shift in rod transcriptome from P6 to P10, consistent with the time of morphological differentiation. We have also identified novel NRL-regulated transcripts generated by alternate splicing or promoter usage. In addition, we discovered a large number of novel photoreceptor transcripts that are regulated by NRL. Our studies have unraveled unappreciated substructures of the NRL-centered gene regulatory network and provided novel secondary regulatory regulators. Noncoding Transcriptome of Developing Rods: By mining our RNA-seq data, together with de-novo assembly, we have identified several known and novel long noncoding RNAs (lncRNAs), which show dynamic expression during photoreceptor development. We have confirmed photoreceptor-enriched expression of 11 novel lncRNAs by high resolution in situ hybridization. Weighted Gene Co-expression Network Analysis (WGCNA) analysis of lncRNAs and associated protein-coding genes and predicted their function may provide new paradigms to elucidate the role of lncRNAs in developing rods. Epigenetic Regulation of Rod Development: To delineate the role of epigenomic modifications, we have produced global profiles of H3K27Ac, H3K4me1 (active/poised enhancers or promoters), H3K36me3 (actively transcribed regions), and H3K9me3 (heterochromatin) marks and ATAC-seq (open chromatin) using P2, P10 and P28 rods. Initial analysis has defined 15 unique chromatin states and overrepresented transitions between chromatin states. Overlapping the known TF binding motif database over chromatin accessibility landscape (obtained by ATAC-seq) has begun to uncover novel regulatory interactions for experimental validations. Evolution of Rod Dominance in Mammals: Our analyses of the mouse rod transcriptome and epigenome revealed a molecular footprint of S-cones, leading us to make a conceptual advance in evolution of rod dominance in mammals (Kim et al. 2016). We have now published a possible mechanism to explain how early mammals may have exploited scotopic niches during nocturnal bottleneck by acquiring rod-specific regulation of NRL. Biogenesis of Photoreceptor Cilium The outer segment of PRs is a modified primary (sensory) cilium that is uniquely designed to capture light. Defects in biogenesis or function of primary cilium can result in pleiotropic clinical findings (ciliopathies), with photoreceptor degeneration being a highly penetrant phenotype. Our focus has been on CEP290, which is a major cause of LCA and is associated with numerous ciliopathies. We have shown that CP110, a major CEP290 interacting protein, is involved in both suppression and promoter of ciliogenesis in vivo (Yadav et al. 2016). Determinants of Photoreceptor Pre-Synapse Morphology Integration of photoreceptors in retinal architecture is essential for visual function. To develop a viable photoreceptor replacement therapy for retinal diseases, we must understand how photoreceptors make specific synaptic connections with inner retinal neurons. Though rods and cones originate from common precursors and perform similar functions, rods have small spherule like presynaptic structure (generally with a single ribbon) connecting to interneurons close to the ONL whereas cones have much larger pedicle with multiple ribbons and telodendrites and connect closer to interneurons. We took advantage of the pedicle-like morphology of the terminals in Nrl-/- retina to identify over 600 potential candidate genes that may be associated with determination of presynaptic structures. We have performed a genetic loss of function screen in rods for >10% of the genes to evaluate their impact on rod pre-synapse morphology. Of these, 18 genes affected the position of terminal within the OPL or the size of the spherule. Further studies are in progress to decipher how structural perturbations affect rod functionality. Cone-Rod Communication and Cone Survival Loss of rods eventually results in the death of cones, which are primary mediators of human vision. In addition to trophic factors such as RDCVF, loss of rod-cone communication (e.g., via gap junctions) and spatial considerations may also be critical for cone survival. Notably, several rows of cones survive and function in Nrl-/- retina after an initial period of cell death, suggesting complex biological paradigms underlying cone homeostasis. We took advantage of our transcriptome data and identified 1046 genes that are enriched after P6 in Nrl-/- S-cones compared to rods and encode membrane or secreted proteins. We are initiating a loss and gain of function screen for 19 selected genes in WT and Nrl-/- mouse retina, with a goal to identify cone genes associated with cell-cell interaction and survival. Rod Differentiation in 3D Retinal Organoids from Mouse ES and iPS Cells Self-organizing neural retina-like structures with proper polarity, lamination and distinct cell types can be generated from mouse and human PSCs in three-dimensional (3D) organoid cultures. We improved the current protocol using hypoxic environment to facilitate efficient development of neural retina from PSCs (Chen et al. 2016). We showed sequential birth of all cell types and longer survival of retinal organoids, mimicking in vivo retinal differentiation and lamination. The transcriptome of rods purified from D35 organoids revealed their maturation just beyond P6 rods in vivo. Our in vitro organoid culture system will be integrated with bioengineering platforms to expedite studies on disease modeling and treatment development (see below). As Muller glia can generate distinct neuronal subtypes upon injury in zebrafish, we examined gene profiles of activated Muller glia after light damage to identify early pathways associated with regeneration response (Sifuentes et al. 2016). Bioengineering Platforms for Development of Mature Neural Retina in vitro Our long-term goal is reconstruction of stem cell-derived retina in vitro for disease modeling and development of therapies. Our approach involves the use of NASA-developed rotating wall bioreactors to enhance nutrient and oxygen supply, and biomimetic scaffolds to provide physical support and mimic retinal microenvironment. We have observed faster differentiation and better maintenance of all cell types in mouse retinal organoids in bioreactor. We are optimizing in vitro culture and differentiation of retinal cells, including photoreceptors, in ECM-mimicking polysaccharide-based hydrogels. Molecular Insights into Human Retina Development The presence of cone-only central fovea and cone-rich macula in the human retina provide high visual acuity and are spatially distinct from the peripheral retina in architecture and function. To elucidate the development of unique spatial and functional organization of the human retina, we have performed RNA-seq analyses of fetal retinas from 52-150 days (d) post-conception and corroborated the findings by immunohistochemistry Our studies show that neurogenesis in the fovea is complete by 96d and is developmentally ahead of the peripheral retina by several weeks. We are also compiling molecular signatures at distinct fetal retinal